Non-Linear Regression

How would we be able to find the line of best fit for data that is supposed to be curved? This is non-linear regression!
In most cases we dont know what the value of a certain point would be on a line of best fit, so we have to guess! It becomes an estimate. Modern technology can help us out a lot by finding the info for us. It even provides us with r2 which is how closely any type of regression curve fits the data.

yy

r2 is the coefficient of Determination defined as

Variation in y explained by variation in x = r2 Total variation in y

( yest y ) r 2 ( y y)
2

is the mean y value

is the y value estimated by the best-fit curve for a given x value

is the actual observed value for a given x value

If the curve is a perfect fit, yest and y will be identical for each value of x, and r2 = 1. If the curve is a poor fit, the total of will be smaller than the total of , therefore r2 will be close to 0. The curve of best fit will be the one that has the highest value for r2. r2 can only have values from 0 to 1.

Possible Non-Linear Regression:

Power or Polynomial Regression uses the line of y = axb as the line that models the information.

Exponential Regression uses the line of y = abx or y = aekx , where e = 2.718 28 , an irrational number commonly used as the base for exponents.

There are limits to the regression curves. It can give inaccurate results to your data. We get a different answer if we use a linear regression. Polynomial regression could give us the wrong info if we perceive it to be exponential! Your model should show a logical relationship between the variables with the best fit possible.
For example: Depreciation of the value of a car exponential Trajectory of a rocket power (quadratic)